SE503644C2 - Ways to determine the content of organic material in effluents from pulp and paper mills - Google Patents

Abstract

PCT No. PCT/EP95/04030 Sec. 371 Date Apr. 11, 1997 Sec. 102(e) Date Apr. 11, 1997 PCT Filed Oct. 10, 1995 PCT Pub. No. WO96/12183 PCT Pub. Date Apr. 25, 1996A method, based on UV/VIS/NIR/IR (including Raman spectroscopy), is provided for the qualitative and quantitative determination of quality parameters in pulp and paper and/or the organic content in effluents from pulp and paper production, by applying chemometric methods.

Description

sus 644 _ 2 The bleached pulp is characterized e.g. of residual lignin content measured as kappa number; paper whiteness measured as ISO-90 or similar parameters; the strength of the paper measured as viscosity, tear index, tensile index or burst index.

The measurement of the kappa number in the aqueous phase is a relatively rough measure of the reduction in lignin content after each step. According to the Technical Association of Pulp and Paper Industries (TAPPI) "the kappa number reflects the relative hardness, bleachability or degree of delignification. The kappa number is defined as the degree of consumption of potassium permanganate (TAPPI Standard Method T 236 cm-85). Another, very similar and standardized method is SCAN-C 1:77.

Viscosity is a surrogate parameter that estimates the strength of the pulp and is often used instead of the slower determination of the strength of the paper through tear index, tensile index and burst index.

Another important quality parameter is the whiteness or lightness of the pulp and paper, which is generally measured as ISO-90.

The impact of the environment on effluents from pulp and paper mills is often assessed through parameters that estimate the organic content. Such parameters are chemical oxygen demand (COD), biological oxygen demand (BOD 5, BOD 7 and BOD 28), the total content of organic matter (TOC), the color of the effluent, adsorbable organohalogens (AOX), suspended organic matter and washing losses, the latter being expressed as COD or sodium sulfate.

A significant disadvantage with the methods known in the prior art for checking the efficiency of the bleaching process is the delay between the sampling of the effluent from the bleaching plant and the determination of the kappa number.

This delay can lead to significant quality losses * - 503 644 3 because the appropriate measurements are unnecessarily delayed. Another disadvantage of the prior art methods is that the characteristics and distribution of this chemically very complex organic material are not well described by these simple methods. It is obvious that there is a definite need for more appropriate and relevant methods for characterizing the quality of the pulp and the organic material in pulp mill effluents. The object of the present invention is to provide a solution to the above-mentioned and similar problems by providing a method which enables continuous control of pulp mill effluents during the pulp and paper milling process. This object is achieved with the combined use of spectrometric and chemometric techniques.

According to the invention, the aqueous effluent is subjected to spectrometric analysis. The water-soluble organic material constitutes a multi-component system or a system with a high degree of background disturbance which greatly complicates the problem of spectrometric analysis.

The use of multivariate data analysis in the characterization of multicomponent systems is currently an area under development. Generally applied in the field of chemistry and especially the field of analytical chemistry, these statistical methods are also called chemometric methods and form the discipline of chemometry. The technique of chemometry is described in more detail in S.D. Brown, "Chemometrics", Anal. Chem. 62, 84R-10lR (1990), which is incorporated herein by reference in its entirety.

Brown et al. In U.S. Patent 5,121,337 (1990) disclose a method based on multivariate data analysis for correcting spectral data with data depending on the spectral measurement process itself and judging unknown properties and / or composition data. in a sample using such a method.

Richardson et al describe in US 5,242,602 (1992) a method for simultaneous analysis of the concentration of "performance indicators" so: 644 - 4 in water systems by determining an absorbance or emission spectrum in a wavelength range of 200 to 2500 nm and application of chemometric algorithms to the absorbance or emission spectrum The said performance indicators are selected from scale inhibitors, corrosion inhibitors, lubricants, metalworking fluids, pH adjusting agents, dispersants, antifoams, chelating agents, biocides, tackifiers and precipitants. The concentrations of "performance indicators" determined by chemical algorithms are compared with predefined areas for the respective "performance indicators".

However, the combination of simultaneous measurement of multiple wavelengths in a fluid sample and application of chemometric techniques has been previously described and used as a general technique.

For example, Hewlett-Packard, a commercial company that manufactures chemical analytical instruments, discusses the importance of chemometric methods for multicomponent analysis using multi-wavelength spectroscopy in its brochure "The diode-array advantage in UV / Visible spectroscopes" (Hewlett-Packard Co., publication). No. 12-5954-8912, 1988).

However, none of the above authors suggests how to solve the problem of qualitative and quantitative determinations of the various parameters relating to pulp and paper production such as the lignin content estimated in the kappa number; the parameters that estimate the strength of the paper, such as viscosity, tensile index or burst index; brightness measured as ISO-90; or the parameters that estimate the organic content of the effluents, such as chemical oxygen demand (COD), washing losses expressed as COD or sodium sulphate, the total organic content (TOC); and biological oxygen demand (BOD 5, BOD 7 and BOD 28); paint, suspended solids, adsorbable organohalogens (AOX), washing losses expressed as COD or sodium sulphate in a way that allows continuous control of these parameters. However, this is the object of the present invention, which provides a reliable and accurate way to control the content of organic material or the parameters which estimate the strength or standard of paper by a rapid chemical analysis coupled with multivariate data analysis. using chemometric techniques.

An object of the invention is thus to provide a method for assessing the differences in kappa numbers in pulp mill effluents before and after a process step by measuring the organic content in the aqueous phase.

Another object of the invention is to provide a method for determining the differences in paper strength measured as viscosity value, tear index, tensile index or burst index, before or after a process step by measuring the organic content in the aqueous phase in pulp mill effluents. Yet another object is the assessment of the differences in the whiteness or brightness of the paper measured as ISO-90 or similar parameters before or after a process step by measuring the organic content in the aqueous phase.

The invention also includes methods for the immediate determination of chemical oxygen demand (COD), washing losses defined as COD or sodium sulphate, total organic carbon content (TOC), biological oxygen demand, e.g. BOD 5, BOD 7 and BOD 28 as well as adsorbable organic halogens (AOX), suspended solids and water-phase paint in the pulp mill effluents.

Summary of the Invention The above object of the invention is obtained by a method for measuring organic material in the aqueous phase of the effluent from pulp and paper mills by analyzing the ultraviolet, visible and / or infrared spectrum of the paper / pulp in the process line in a wavelength range from 180 nm to 2400 nm, and applies chemometric evaluation of the spectrum to calculate the differences in kappa numbers; differences in paper strength measured as viscosity, tear index, tensile index, burst index or similar parameters; the difference in the whiteness or brightness of the paper measured as ISO-90 or similar parameters, before and after a process step. Further finishes are the direct determination of chemical oxygen demand (COD), washing losses expressed as COD or sodium sulphate, total organic content (TOC), as well as biological oxygen demand (BOD 5, BOD 7 and BOD 28), adsorbable organic halogens (AOX), suspended solids material and color.

Detailed description of the invention According to the invention, it has now been shown through extensive development work that it is possible to directly and continuously detect absorption or transmittance spectra of the parameters defined above and especially the kappa number in the pulp and / or COD in the pulp mill effluent. a UV / VIS-NIR spectrometer and, by using the absorbance or transmittance values at discrete wavelengths from these spectra, calculate the parameters defined above, in particular the kappa number.

The term pulp as used herein refers not only to bleached pulp and / or paper but also to unbleached or partially bleached mechanical or chemical pulp.

Technically, the spectrometric analysis can be performed by on-line, in-line or at-line detection and can be performed as a continuous check using an on-line, in-line or at-line optical fiber probe or by to take individual samples for separate analysis. In both cases, absorption or transmittance spectra are subjected to further data processing using values from several discrete wavelengths from each particular spectrum.

The spectral information estimates a variety of properties in the organic content. Depending on the parameters that are of interest, relevant and selected spectral information is correlated to the specific parameter.

An example of such a technique is the use of a device, located at a distance from the process, containing a light source, detector, electronic components and other components necessary to transmit a signal through an optical fiber to the sample, where the light is transmitted through or estimated on or in part through the sample. The resulting signals are fed back to the detector via an accompanying optical fiber cable, and recorded.

In the spectrometer, the light is transmitted to an electrical signal which is then transmitted to a computer where the spectrum of a previously stored reference scan can be related to, e.g. subtracted from, the sample spectrum and a reference-corrected spectrum are calculated.

Another example is by manually or automatically taking samples at relevant time intervals and subjecting the samples to analysis in an analytical instrument containing the light source, detector, electronic components and other necessary components. Absorption or transmittance spectra are then subjected to further data processing using values from several discrete wavelengths from each particular spectrum.

It is preferred that the detector have a measuring range of at most 10 nm, preferably 2 nm, and most preferably 1 nm or less.

The detection is performed in the UV-VIS-NIR wavelength range of 180 nm to 2400 nm.

This can be done using a scanning instrument, a diode vector instrument, a Fourier transform instrument or any similar equipment known to one skilled in the art.

An evaluation of wavelengths that contain absorption or transmission provides information relevant to the analysis. Through sus 644 - 8 application of chemometric methods to obtained spectra, it is then possible to ignore wavelengths that do not contain information that contributes to the chemical analysis, even if the measurement includes information from the entire wavelength range.

The determination and control of e.g. the reduction in kappa number of pulp and / or paper using the spectrometric measurements comprises two main steps, the first step being the development of a calibration model which includes the development of the reference set; data processing; and data analysis, through the use of pulp and / or paper samples with known kappa numbers. The second main step is the spectrometric analysis of the sample with unknown kappa number, spectral data processing, possibly as a result of data analysis; and applying the calibration model, developed in the first main step, to the data obtained thereby.

(I) DEVELOPMENT OF A CALIBRATION MODEL The cap number is measured in the traditional way (TAPPI T 236 cm-85 (1984) or SCAN-C 1:77) for a number of mass samples before and after a process step. The difference in kappa number, before and after the process step, was then used in the development of a calibration model in which the three sub-steps discussed below are applied to registered absorption, reflectance or emission spectra for the said sample.

(I.a) Development of the reference set The model reference set consists of a large number of absorption or transmission spectra from samples with known kappatal characteristics, which should preferably be representative of the production line. The reference set was used in the chemometric algorithm to calculate the resulting model parameters.

(I.b) Data processing To reduce noise and adjust for baseline operation, spectral raw data should be processed. This processing can also reveal hidden information, such as identity with obviously different spectra or non-identity of obviously very similar spectra.

In addition, the assumptions leading to Beers' law (which states that for a given absorption coefficient and length of an optical path in the absorptive medium, the total amount of absorbed light is proportional to the sample's molecular concentration) are seldom met in the complex system of pulp mill effluent samples. This is due to a number of factors, such as high concentrations, which are often found in industrial samples. Another complicating factor is light scattering variations due to particles in the sample.

Various theories have been developed to overcome this problem and the most widely used is: the Kubelka-Munk transformation (P.

Kubelka, F. Munk, Z. Tech. Physik 12, 593 (1931)), which involves absorption and scattering; and "the Multiplicative Scatter Correction" (P. Geladi, D. MacDougall, H. Martens, Appl. Spect. 39, 491-500 (1985)) where each spectrum is "corrected" in both offset and slope by comparison with a "ideal" spectrum (medium spectrum).

The Kubelka-Munk transformation is according to equation 1: = (l- ¿k) 2 (1) Aik zR-ik where Rik is the apparent absorbance at wavelength k. Aik is the transformed absorbance at wavelength k and the index i shows available sample spectra.

"The Multiplicative Scatter Correction" (MSC) is according to Equation 2: (2) sus 64-4 _. Wherein Aik, Rik, i and k have the same meaning as above, âi is the least squares calculation of the intercept parameter and bi is the least squares calculation of the slope parameter.

Another way of linearizing spectral data is the use of derivatives, e.g. up to a fourth degree derivative (A. Savitzky, M.J.E. Golay, Anal. Chem. 36, 1627-1639 (1964)). The derivatives of the spectrum result in a transformed spectrum, consisting only of the relative changes between the adjacent wavelengths and it has been shown that the peak intensities of the derived spectra appear to be more linear with the concentration (TC O'Haver, T. Begley, Anal. Chem. 53, 1876 (1981)).

Linearization can also be accomplished using the Fourierer transform or using the Standard Normal Variate Transformation described by R.J. Barnes, M.S. Dhanoa and S.J. Lists in Appl. Spectrosc., Volume 43, number 5, p. 772-777 (1989).

(I.c) Data analysis Data analysis using chemometric techniques then allows the development of the calibration model. There are several chemometric techniques that can be used, such as principal component analysis (PCA), partial least square regression (PLS), principal component regression (PCR), multilinear regression analysis (MLR) and discriminant analysis. The preferred chemometric technique of the invention is the PLC method.

(I.c.1) Principal Component Analysis (PCA) PCA reduces a set of correlated variables to a smaller number of uncorrelated variables. This transformation consists of a rotation of the coordinate system, which results in the centering of information on a smaller number of axes than in the original coordinate system. This treats variables that are highly correlated with each other as a single unit. Thus, with the use of PCA, it is possible to obtain a 3 ~ 505,644 ll smaller set of uncorrelated variables which still represents most of the information present in the original set of variables but which is much easier to use in models. In general, 2 to 15 principal components give 85 to 98% of the variance of the variables.

(I.c.2) Partial least square regression (PLS) PLS is a modeling and calculation method by which quantitative relationships can be established between blocks of variables, e.g. a block of descriptor data (spectrum) for a series of samples and a block of response data measured on these samples. Due to the quantitative relationship between the blocks, it is possible to obtain spectral data for a new sample for the descriptor block and make predictions of the expected answers. A major advantage of the method is that the results can be evaluated graphically through different curves. In most cases, a visual interpretation of the curve is sufficient to gain a good understanding of the different relationships between the variables. The method is based on projections, similar to PCA. The PLC method is described in detail in Carlsson R., Design and optimization in organic synthesis, B.G.M. Vandeginste, O.M. Kvalheim, ed., Data handling in science and technology, (Elsevier, 1992), vol. 8.

(I.c.3) Principle Component Regression (PCR) PCR is closely related to PCA and PLC. As in PLC, each object in the descriptor block is projected onto a space with fewer dimensions than the original, which results in the information being able to be characterized with so-called "scores" and "loadings".

Via a regression procedure, the response matrix is modeled against the spectral information, expressed as "scores". This can be used to predict unknown samples. The same model statistics as in PLC and PCA can be used to validate the model.

For an excellent tutorial on PCA, PLC and PCR, see P. Geladi et al in "Partial Least-Squares Regression: A Tutorial" in Anal.

Chim. Acta, 185, l-32 (1986), which is incorporated by reference in its entirety sus 644 - 12 herein by reference.

(I.c.4) Multilinear regression analysis §MLR) MLR defines the best fit plane for the kappa number as a function of the spectrum using the least squares method to define each boundary line of the plane.

This plan is then used to recognize and assign a predicted value for an unknown kappa number. This technique is generally limited to relatively "pure" systems where there is no significant amount of matrix interference and unlike PLC it requires more objects than variables.

(I.c.5) Discriminant analysis This is a method whereby, by using spectral data, the values of the known kappa numbers are grouped into different clusters, separated by linear determination limits. Through its spectrum, a sample with unknown kappa number can be matched against a cluster and the kappa number can be assigned a value, e.g. the mean of the cluster. This is a very useful technology for quality screening but requires a much larger database to get statistically significant results.

(II) DETERMINATION OF ORGANIC MATERIAL BY APPLICATION OF THE CALIBRATION MODEL Once a calibration model has been developed, the unknown kappa number can be determined by recording the absorption or transmission spectrum in accordance with (I.a) from the pulp mill effluent with unknown kappa; processing of thereby obtained spectral raw data according to (I.b); optionally performing data analysis on processed spectral data according to (I.c); and application of the developed calibration model to the data thus obtained.

The invention will now be illustrated by an example. å - sus 644 13 EXAMPLES Multi-wavelength spectroscopy, performed on effluents from mass bleaching, followed by linearization of spectral data and multivariate data determination (PLS algorithm) was used to determine the reduction in kappa number before and after bleaching. The organic content is expressed as kappa number reduction between two bleaching steps.

DEVELOPMENT OF A CALIBRATION MODEL (A) Development of the reference set BBQ! The reference samples consisted of a total of 55 effluents from chemical bleaching compound from different places in the bleaching sequence and thereby with different pH and COD levels. The samples had been treated in different bleaching sequences and the kappa number was determined for each sample using TAPPI standard method T 236 cm-85.

UV ABSORBANCE MEASUREMENTS The UV measurements were obtained using a "Spectra Focus® Optical Scanning Detector" (Spectra-Physics), equipped with a fused silica capillary with an inner diameter of 0.53 mm as flow through the cell. The measurements were made directly on the capillary on which a 10 mm opening had been obtained by burning off the polymer coating. The samples were pumped through the capillary using a peristaltic pump (Gilson Minipulse 2) at a rate of 2.4 ml min'1. A spectral working range between 200 nm and 360 nm, in even intervals of 2 nm, was used and gave 81 measurements at different wavelengths. Four measurements were made on each sample and their mean spectrum was calculated and used in further evaluations. The obtained absorbance data was transferred to an IBM PS-2 7386 computer equipped with AutolaU® Software (Spectra-Physics).

(B) Data processing The entire spectral data matrix was used for modeling, resulting in 81 spectral points. sus 644 - 14 LINEARIZATION TRANSFORMATION Two linearization functions were used, ie. the second derivative and the MSC of the original spectrum. Models were built on both the original data matrix and the transformed one.

(C) Data analysis The commercially available software MATLAB V 4.0 was used for numerical calculations. The PLS algorithm was used to model the relationship between spectra and descriptors as a general function in the commercially available "Chemometrics Toolbox" based on the NIPALS algorithm (H. Wold, P. Krishnaiah, Multivariate Analysis, 391 (1966) ). The method for establishing the significant number of PLC components was crush validation (S. Wold, Technometrics 20, 397-405 (1978)) (Jackknifing) whereby a sample was omitted. This procedure was repeated in an iterative manner, omitting every other sample so that each predicted value was based on a model in which the predicted value was not present. The capital values were centered and scaled to unit variance before modeling (autoscaling or z-transform) and rescaled before model evaluation.

RESULTS.

The measured against the modeled values of the reduction in kappa number are plotted in Fig. 1. As can be seen, the deviation from the theoretical value, represented in the figure as a straight line, is minimal.

In Fig. 1, the values, calculated from the reference method, are plotted against the values calculated by the model (predicted values).

The figure shows the prediction of the kappa number using a PLC model with 7 principal components. Data pretreatment was multiplicative scatter correction and mean centering. The solid line shows slope 1.0, the dashed line was adjusted to linear regression, while the dotted lines show the confidence limit at 95% significant level (SEP = Standard error of prediction: RMSECV = Root mean squared error of cross * - sus 644 15 Validation; MSECV = Mean squared error of cross Validation; INTER = intercept; K = slope; rA2 = the square of the correlation coefficient for the fitted dashed line).

Ideally, the slope of the line (k) and its correlation coefficient (r2), as well as R2Y (cum) and Q2 (cum) should be as close to l as possible, while the intercept should be as close to 0 as possible. With regard to the values obtained, it is possible to realize very good quality and model precision.

The advantage of the new method of controlling the organic content of paper or pulp effluent using chemometry should therefore be obvious. A variety of samples from different pulp mills and production processes can be analyzed using the same calibration set. The invention thus provides a method whereby the control of the quality of the pulp and the organic content of effluents from the production process in pulp and paper mills can take place in a very fast and precise manner on any type of pulp and paper.

Claims (14)

sus 644 _ 16 PATENTKRAV Method for the qualitative and quantitative determination of the organic content of pulp and paper production effluents by (I) development of a calibration model by (Ia) recording absorption or transmission spectra of reference samples with known properties and concentration of organic substances for the development of the reference set; (I.b) processing of spectral raw data to reduce noise and adjust for operation and diffuse light scattering; (I.c) performing data analysis by applying chemometric techniques to the processed reference sets; and thus calibration of processed spectral data with the actual values obtained from analytical reference methods; (II) determining the unknown properties and / or organic substances in pulp and paper by recording absorption or transmission spectra, in accordance with (I.a): processing the spectral raw data thus obtained according to (I.b); and application of the developed calibration model to processed data, knowledge of qualitative and quantitative determination of the parameters that estimate the quality of pulp or paper and the environmental impact of the effluents by measuring the organic content of the aqueous phase obtained in pulp and paper production. Method according to claim 1, characterized in that the difference in product quality before and after a process step is calculated and that this value is qualitatively and quantitatively related to the spectral information obtained by measuring the process effluent. Method according to claim 2, characterized in that the difference in kappa number before and after a bleaching step is calculated and that this value is quantitatively related to the selected part of the spectral information obtained by measuring the bleaching plant effluent. 4. A method according to claim 2, characterized in that the difference in paper strength, measured as a viscosity value before and after a bleaching step, is calculated and that this value is quantitatively related to the selected part of the spectral information obtained by measuring the bleaching effluent. Method according to claim 2, characterized in that the calculation of the difference in paper strength, measured as tear index, tensile index, burst index and similar parameters, is calculated before and after a bleaching step, and that this value is quantitatively related to the selected part of the spectral information obtained by measuring the bleaching plant effluent. Method according to claim 2, characterized in that the difference in the whiteness or brightness of the paper, measured according to ISO-90 or similar parameters, is calculated before and after a bleaching step and that this value is quantitatively related to the selected part of the spectral information obtained by measuring the bleaching plant effluent. Method according to claim 1, characterized in that the organic content of the pulp and paper process effluents is determined and that this value is qualitatively and quantitatively related to selected parts of the spectral information obtained from the process effluent. Method according to claim 7, characterized in that the organic content, defined as chemical oxygen demand (COD), in the aqueous effluent from the production in a pulp mill or a paper mill is determined and that this COD value is quantitatively related to selected parts of the spectral information obtained from the effluent. sus 644 «l8 Method according to claim 7, characterized in that the organic content, defined as total organic carbon content (TOC), in the aqueous effluent from the production from pulp mill or paper mill is determined and that the TOC value is quantitatively related to selected parts of the spectral information obtained from the effluent. 10. A method according to claim 7, characterized in that the organic content, defined as biological oxygen demand (BOD), e.g. BOD 5, BOD 7 and BOD 28, in the aqueous effluent from the production from a pulp mill or paper mill are determined and that the BOD value is quantitatively related to selected parts of the spectral information obtained from the effluent. 11. ll. Method according to claim 7, characterized in that the organic content, defined as adsorbable organic halogenates (AOX), in the aqueous effluent from the pulp mill or paper mill is determined and that the AOX value is quantitatively related to selected parts of the spectral information. obtained from the effluent. Method according to claim 7, characterized in that suspended solids in the aqueous effluent from the production from pulp mill or paper mill are determined and that the value of suspended solids is quantitatively related to selected parts of the spectral information obtained from the effluent . Method according to claim 7, characterized in that the color of the aqueous effluent from the production from pulp mill or paper mill is determined and that the COD value is quantitatively related to selected parts of the spectral information obtained from the effluent. Method according to claim 7, characterized in that washing losses, defined as chemical oxygen demand (50D 644 19 (COD) or sodium sulphate), in the aqueous effluent from production at pulp mill or paper mill are determined and that the washing losses are quantitatively related to selected parts of the spectral information obtained by measuring the effluent.